Is Router One's subscription unlimited?

No — and be wary of anyone who says unlimited. Plans carry monthly quotas by model tier; when you exceed them, overage bills to your wallet balance at posted per-token rates so sessions keep running. Current quotas are listed on /pricing.

Does Claude Code really work without official Claude Max?

Yes. Claude Code talks to any Anthropic-compatible endpoint. Set ANTHROPIC_BASE_URL to https://api.router.one and ANTHROPIC_AUTH_TOKEN to your Router One key — no wrapper or proxy needed. Codex CLI is supported the same way on the OpenAI-compatible endpoint.

How is this cheaper than paying Anthropic or OpenAI directly?

Two mechanisms: subscription plans price a monthly usage quota at a flat rate, and pay-as-you-go rates run up to 90% off official prices on select models. The exact saving depends on your workload — run it through /llm-cost-calculator.

What happens when I hit my plan quota mid-session?

Overflow bills to your wallet balance at posted per-token rates instead of cutting the session off. You can cap exposure with per-key budgets and rate limits.

Which payment methods work?

Card payments, USDT/USDC stablecoins on six networks, Alipay, or WeChat Pay — no US credit card required. Wallet top-ups and subscriptions both work with any of these methods.

Why not just use OpenRouter?

OpenRouter passes through official list prices plus platform fees, which leaves a heavy user's bill essentially unchanged. Router One changes the unit economics with quota subscriptions and discounted select models — see /openrouter-alternative for the full comparison.

A Claude Max alternative built for heavy Claude Code users

Agentic coding burns tokens at a rate consumer subscriptions and pay-per-token billing were never designed for. Official Claude Max runs $100–$200 a month for one vendor's models; raw API billing for a heavy Claude Code user can pass that in a week. Router One subscriptions start at $20/month, meter usage against a monthly quota instead of cutting you off or surprising you, and cover 40+ models — Claude included — through one OpenAI-compatible endpoint.

Get your API key

Where the bill shock comes from

Official subscriptions are per-vendor and capped

Claude Max at $100–$200/month covers Claude only, with usage limits that heavy agent sessions hit mid-task. When you also need GPT or Gemini for other workloads, each vendor wants its own subscription.

Pay-per-token API billing scales with your ambition

A Claude Code power user running long agentic sessions can consume tens of millions of tokens a month. At official API list prices that is hundreds of dollars — sometimes more than an engineer's entire tool budget.

Pass-through resellers don't change the math

Marketplaces that route to official APIs at list-price pass-through (plus platform fees) leave the bill exactly where it was. The unit economics only change with quota-based plans or discounted rates.

How Router One subscriptions work

Flat monthly price: Pro, Max, and Ultra plans (currently $20 / $48 / $149 — see /pricing for live plans and quotas).
Metered monthly quotas by model tier instead of hard mid-task cutoffs; overage bills to your wallet balance at posted per-token rates, so nothing silently stops.
Native Claude Code and Codex CLI support: an Anthropic-compatible base URL for the Claude family plus the OpenAI-compatible endpoint for everything else.
One subscription spans 40+ models (GPT, Claude, Gemini, DeepSeek, Mistral and more) — no per-vendor stacking.
Pay with card, USDT/USDC on six networks, Alipay, or WeChat Pay — and the same endpoint stays reachable worldwide, including mainland China.

Claude Max vs Router One at a glance

	Claude Max (official)	Router One
Monthly price	$100–$200	Plans from $20; pay-as-you-go wallet also available
Models covered	Claude family only	40+ models across GPT, Claude, Gemini, DeepSeek, Mistral
Claude Code	Included, subject to plan limits	Native support via Anthropic-compatible endpoint, metered by plan quota
API keys for your own apps	Not included (separate API billing)	Included — the same account serves CLI tools and your applications
Overage behavior	Wait for the limit window to reset	Overflow bills to wallet balance at posted rates — sessions keep running
Payment methods	International credit card	Cards, USDT/USDC (6 networks), Alipay, WeChat Pay
Mainland China access	VPN required	Direct connection, no VPN
Per-request cost visibility	None	Model, tokens, cost, latency and status logged per request

Point Claude Code at Router One in two minutes

No wrapper, no proxy config — Claude Code speaks to Router One natively. Set two environment variables and keep working:

terminal

export ANTHROPIC_BASE_URL=https://api.router.one
export ANTHROPIC_AUTH_TOKEN=sk-your-router-one-key

claude   # Claude Code now routes through Router One

Run your own numbers

Every workload is different, so we built a calculator instead of a slogan: plug in your monthly request volume and token sizes, and compare official API list prices against Router One pay-as-you-go and subscription plans side by side. For per-model rates, the full catalog with posted prices is public.

Cost calculator: /llm-cost-calculator — your usage, priced three ways.
Model catalog and posted per-token rates: /models.
Current plans and quotas: /pricing.

FAQ

Is Router One's subscription unlimited?: No — and be wary of anyone who says unlimited. Plans carry monthly quotas by model tier; when you exceed them, overage bills to your wallet balance at posted per-token rates so sessions keep running. Current quotas are listed on /pricing.
Does Claude Code really work without official Claude Max?: Yes. Claude Code talks to any Anthropic-compatible endpoint. Set ANTHROPIC_BASE_URL to https://api.router.one and ANTHROPIC_AUTH_TOKEN to your Router One key — no wrapper or proxy needed. Codex CLI is supported the same way on the OpenAI-compatible endpoint.
How is this cheaper than paying Anthropic or OpenAI directly?: Two mechanisms: subscription plans price a monthly usage quota at a flat rate, and pay-as-you-go rates run up to 90% off official prices on select models. The exact saving depends on your workload — run it through /llm-cost-calculator.
What happens when I hit my plan quota mid-session?: Overflow bills to your wallet balance at posted per-token rates instead of cutting the session off. You can cap exposure with per-key budgets and rate limits.
Which payment methods work?: Card payments, USDT/USDC stablecoins on six networks, Alipay, or WeChat Pay — no US credit card required. Wallet top-ups and subscriptions both work with any of these methods.
Why not just use OpenRouter?: OpenRouter passes through official list prices plus platform fees, which leaves a heavy user's bill essentially unchanged. Router One changes the unit economics with quota subscriptions and discounted select models — see /openrouter-alternative for the full comparison.

Get your API key